What were the problem faced?

What were the challenges?

The client previously had engaged with another offshore development partner since the inception of the product. The code of the product had evolved to be clunky and grown difficult to maintain. Certain parts of the system were not scalable and were an operational overhead. Debugging production issues was very difficult.

The client tried to improve the process and overall product quality with help of internal people along with the external development partner which did not result in much success.

Engineering for an Expense Report Auditing product

UI code which was written in AngularJS with heavy usage of jQuery. Both did not go well together and resulted in code with thousands of lines of unreadable code.

The customer had a module for running scheduled jobs. This module had the ability to run schedules and run multiple jobs per customer. There were occasions when jobs did not fire or froze while running. Hung jobs caused database deadlock issues at times.

There was no way to track what a particular job was doing at a time.

The client was unable to roll out new features faster and a large part of the time was spent in managing production issues. The client was finding it difficult to onboard new customers due to delays in development. Inability to onboard a customer was a straight loss of business.

Proposed Solution

Our team rebuilt and restructured many modules. We cleaned up the UI code, removed jQuery and made it purely AngularJS which reduced the size of code by 50% and improved readability and maintainability significantly.

We redesigned the job scheduler module and added following features:

• Created a basic framework around jobs.
• The framework registers each job run within the database
• The framework provides each job run with a unique id and includes the unique job id in the log files to improve traceability of jobs.
• The framework created around jobs abstracted out the operations and reporting aspects, leaving the job developers to only focus on business logic.
• Created alerts emails for events like a stuck job or a failed job.
• We redesigned the job scheduler to be clustered. The jobs could run on multiple machines instead of single machines making it more scalable.
• The redesign allowed the ops team to quickly identify issues and resolve the customer complaints in a timely manner. Ops overhead was reduced by 25%.

How did the client benefit?

• The UI code was significantly reduced and became more readable and maintainable.
• The job scheduler was made a lot more scalable.
• Debugging and troubleshooting became extremely easy for the Ops team and they could react to customer issues much faster.
• Frequent release cycles allowed clients to roll out fixes and new features more frequently.

• The client could onboard a lot more customers due to integrations with third-party systems.